Methodological enchancement of hedonic price index in Malaysia

Housing market is one of the largest asset sectors in Malaysia. Several parties are concerned about the performance of housing market including government, financial institution, market player and public. A proper computation of house price index using hedonic model is critical in monitoring housing performance to avoid biased estimation which due to complexity and misspecification of the model. This study has enhanced the conventional hedonic price index model through a combination of regression modelling and spatial model in order to yield more reliable house price index. The studied sample comprises 6,420 transactions of double storey terrace houses in Johor Bahru from year 2006 until year 2011. This study examined three types of regression modelling namely shrinkage, semiparametric and ordinary least squares method in constructing spatial hedonic price index model and conventional hedonic price index model respectively. An optimal hedonic price index model was ascertained according to the predictive power, accuracy and consistency test in the study. The result found that shrinkage estimator is robust when it comes to perform spatial hedonic price index model as compared to ordinary least squares and semiparametric method. Moreover, the house price index is further enhanced using temporal aggregation and seasonality analysis. The results show that seasonal adjusted monthly index is more effective in monitoring housing price performance. Therefore, shrinkage estimator, spatial hedonic model, temporal aggregation and seasonality analysis are important in enhancing the methodological aspects in constructing hedonic price index. In conclusion, the improved house price index can be used in formulating more effective housing policies and investment strategies.